PDE based Segmentation of Vector-valued Texture Images using Sobolev GradientsCP2

In this paper, we propose a method for minimization of segmentation model for vector-valued texture images. The texture in the image will be smooth by using $L_0$ gradient norm and then the segmentation model will be minimized through sobolev gradient for fast convergence. The better performance of the method will observed from the experimental results. Results of the proposed method are compared with $L^2$ gradients.

This presentation is part of Contributed Presentation “CP2 - Contributed session 2

Fahim Ullah (University of Engineering and Technology Peshawar)
Noor Badshah (University of Engineering and Technology Peshawar)
Hassan Shah (University of Engineering and Technology Peshawar)
image segmentation, level sets, l_0 norm, partial differential equation models, pdes, semi-implicit method., sobolev gradients, vector-valued texture images